Download Tapanes-Castillo A, Weaver EJ, Smith RP, Kamei Y, Caspary T, Hamilton-Nelson KL, Slifer SH, Martin ER, Bixby JL, Lemmon VP. Neurogenetics. 2012 Feb;11(1):53-71. A modifier locus on chromosome 5 contributes to L1 cell adhesion molecule X-linked hydrocephalus in mice.

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Dominance (genetics) wikipedia , lookup

Epigenetics of neurodegenerative diseases wikipedia , lookup

Mutagen wikipedia , lookup

Biology and sexual orientation wikipedia , lookup

Point mutation wikipedia , lookup

Polycomb Group Proteins and Cancer wikipedia , lookup

Chromosome wikipedia , lookup

Gene expression profiling wikipedia , lookup

Designer baby wikipedia , lookup

RNA-Seq wikipedia , lookup

Gene expression programming wikipedia , lookup

Hardy–Weinberg principle wikipedia , lookup

Artificial gene synthesis wikipedia , lookup

Genomic imprinting wikipedia , lookup

Epistasis wikipedia , lookup

Saethre–Chotzen syndrome wikipedia , lookup

Epigenetics of human development wikipedia , lookup

Microevolution wikipedia , lookup

Site-specific recombinase technology wikipedia , lookup

Polyploid wikipedia , lookup

History of genetic engineering wikipedia , lookup

Nutriepigenomics wikipedia , lookup

Y chromosome wikipedia , lookup

Mir-92 microRNA precursor family wikipedia , lookup

Genome (book) wikipedia , lookup

Karyotype wikipedia , lookup

Tag SNP wikipedia , lookup

Neocentromere wikipedia , lookup

Epigenetics in learning and memory wikipedia , lookup

Genome-wide association study wikipedia , lookup

Public health genomics wikipedia , lookup

Skewed X-inactivation wikipedia , lookup

NEDD9 wikipedia , lookup

X-inactivation wikipedia , lookup

Quantitative trait locus wikipedia , lookup

Transcript
Neurogenetics
DOI 10.1007/s10048-009-0203-3
ORIGINAL ARTICLE
A modifier locus on chromosome 5 contributes to L1 cell
adhesion molecule X-linked hydrocephalus in mice
Alexis Tapanes-Castillo & Eli J. Weaver & Robin P. Smith & Yoshimasa Kamei &
Tamara Caspary & Kara L. Hamilton-Nelson & Susan H. Slifer & Eden R. Martin &
John L. Bixby & Vance P. Lemmon
Received: 6 April 2009 / Accepted: 8 June 2009
# Springer-Verlag 2009
Abstract Humans with L1 cell adhesion molecule (L1CAM)
mutations exhibit X-linked hydrocephalus, as well as other
severe neurological disorders. L1-6D mutant mice, which are
homozygous for a deletion that removes the sixth
immunoglobulin-like domain of L1cam, seldom display
hydrocephalus on the 129/Sv background. However, the
same L1-6D mutation produces severe hydrocephalus on the
C57BL/6J background. To begin to understand how
L1cam deficiencies result in hydrocephalus and to identify
modifier loci that contribute to X-linked hydrocephalus by
genetically interacting with L1cam, we conducted a
genome-wide scan on F2 L1-6D mice, bred from L1-6D
129S2/SvPasCrlf and C57BL/6J mice. Linkage studies,
utilizing chi-square tests and quantitative trait loci mapping techniques, were performed. Candidate modifier loci
were further investigated in an extension study. Linkage
was confirmed for a locus on chromosome 5, which we
named L1cam hydrocephalus modifier 1 (L1hydro1),
p ¼ 4:04 " 10#11 .
Keywords L1cam . Hydrocephalus . Modifier . Linkage
analysis . QTL
Electronic supplementary material The online version of this
article (doi:10.1007/s10048-009-0203-3) contains supplementary
material, which is available to authorized users.
A. Tapanes-Castillo : E. J. Weaver : R. P. Smith : J. L. Bixby :
V. P. Lemmon (*)
The Miami Project to Cure Paralysis,
Miller School of Medicine, University of Miami,
Lois Pope LIFE Center, Room 4-16, 1095 NW 14th Terrace,
Miami, FL 33136, USA
e-mail: [email protected]
E. J. Weaver : Y. Kamei : V. P. Lemmon
Department of Neuroscience, Case Western Reserve University,
Cleveland, OH, USA
R. P. Smith : J. L. Bixby : V. P. Lemmon
Neuroscience Program, University of Miami,
Miami, FL, USA
T. Caspary
Department of Human Genetics, Emory University,
Atlanta, GA, USA
K. L. Hamilton-Nelson : S. H. Slifer : E. R. Martin
Dr. John T. MacDonald Foundation,
Department of Human Genetics,
Miami Institute for Human Genomics, University of Miami,
Miami, FL, USA
J. L. Bixby : V. P. Lemmon
Department of Neurological Surgery, Miller School of Medicine,
University of Miami,
Miami, FL, USA
J. L. Bixby
Department of Molecular and Cellular Pharmacology,
Miller School of Medicine, University of Miami,
Miami, FL, USA
Present Address:
Y. Kamei
Department of Maternal, Fetal and Neonatal Medicine,
University of Tokyo Hospital,
Tokyo, Japan
Neurogenetics
Introduction
Hydrocephalus is a devastating neurological condition,
characterized by the abnormal accumulation of cerebrospinal fluid (CSF) within the cerebral ventricles. Ineffective treatment can lead to cerebral atrophy and death.
Human hydrocephalus is categorized based on two criteria:
(1) time of onset and (2) type of CSF flow defect.
Depending on the time of onset, hydrocephalus is classified
as congenital or acquired. Congenital hydrocephalus arises
prenatally, and it occurs with an estimated incidence of 1 in
1,500 births [1, 2]. Acquired hydrocephalus, which
develops during or after birth, may be caused by infectious
diseases, traumatic brain injuries, or tumors. In addition,
hydrocephalus is classified as communicating (nonobstructive) or non-communicating (obstructive) based
on the reason for the defect in CSF flow. Communicating
hydrocephalus arises from problems with CSF secretion or
resorption; CSF flows freely between the ventricles and
the subarachnoid space. On the other hand, noncommunicating hydrocephalus develops when a physical
obstruction blocks CSF flow between the cerebral ventricles and the subarachnoid space.
Linkage analysis of numerous human kindreds with
congenital hydrocephalus led to the discovery that mutations in the L1 cell adhesion molecule (L1CAM) gene are
responsible for X-linked hydrocephalus [3], a form of
hydrocephalus that has been estimated to account for
approximately 7–15% of all congenital hydrocephalus cases
[4]. While the autosomal locus 8q12.2-21.2 has also been
associated with hydrocephalus [5], additional genes have
not been identified. Hence, the molecular etiology of
human hydrocephalus is poorly understood.
Fortunately, the use of animal models has provided
some insight into the genetics of congenital hydrocephalus. Studies on mutant mice have identified at least 14
genes that cause hydrocephalus when disrupted, including L1cam, metal response transcription factor 2 (Mtf2),
platelet-activating factor acetylhydrolase (Pafah1b1/
Lis1), hydrocephalus with hop gait (hyh), and nuclear
factor I-A ([6–12], reviewed by [13]). Overexpression of
transforming growth factor-β1 has also been shown to
cause hydrocephalus in mice [14, 15]. In addition, a
quantitative trait loci (QTL) study in mice revealed three
loci that regulate cerebral ventricle size [16], and QTL
analyses of the Hydrocephalic Texas strain HTX rat led to
the mapping of four QTLs that contribute to hydrocephalus [17]. Furthermore, mouse studies demonstrating that
hydrocephalus can be caused by mutations in ciliary genes
such as Hydin and Dynein axonemal heavy chain 5
(Mdnah5) [18–20] have been complemented by data
generated from zebrafish. Three genes involved in ciliary
function have been associated with the development of
hydrocephalus in zebrafish: polycystin 2, fleer, and oralfacial-digital type 1 syndrome gene (ofd-1) [21–23].
L1CAM, the only gene currently linked to human
hydrocephalus, encodes a Type 1 transmembrane protein
that belongs to the immunoglobulin (Ig) superfamily of cell
adhesion molecules. The L1 protein consists of six
extracellular Ig-like domains, followed by five extracellular
fibronectin type III domains, a single transmembrane
region, and a cytoplasmic tail. Neuronal L1 is expressed
on the surface of axons and growth cones both in the
central nervous system (CNS) and in the peripheral nervous
system. Schwann cells and other non-neuronal cells,
including melanocytes and lymphocytes, also express L1
[24]. L1 mediates cell–cell adhesion, as well as the
adhesion of axons to adjacent axons and the extracellular
matrix (reviewed in [25]). L1 regulates axon outgrowth
(extension and branching), fasciculation, myelination, and
neuronal migration [26–31]. L1 has additionally been
implicated in synaptic function [32–34].
L1 can function through homo- (L1-L1) or heterophilic
interactions [35]. A number of heterophilic binding partners
for L1 have been identified, including TAG-1/axonin-1, F3/
F11/contactin, phosphacan, neurocan, neuropilin-1, and various integrins (reviewed in [25, 36]). The fibroblast growth
factor receptor (FGFR) has also been suggested to directly
interact with L1 [37, 38]. Intracellularly, L1 has been shown
to signal through Src and mitogen-activated protein kinase 1
(MAPK1/ERK2) and to interact with adaptor molecules such
as ezrin, ankyrin, and Ran-binding protein M [39–46].
In humans, mutations in L1CAM are responsible for Xlinked hydrocephalus, agenesis of the corpus callosum,
corticospinal tract hypoplasia, and MASA syndrome—
mental retardation, aphasia, spastic paraplegia, and adducted
thumbs. A comparison of case studies and genetic mutations
revealed a striking correlation between the severity of
neurological disease and the type of mutation harbored by
the L1CAM gene [47–49]. Mutations that generate truncations in the extracellular domain of L1 are more likely to be
lethal or to produce severe hydrocephalus and grave mental
retardation than point mutations in the extracellular domain
or mutations that solely affect the cytoplasmic domain.
Furthermore, point mutations in the extracellular domain
tend to cause more severe neurological problems than
cytoplasmic domain mutations.
The severity of hydrocephalus seen in individuals with
L1CAM mutations is highly variable. Patients range from
displaying no hydrocephalus to those having high-pressure
progressive hydrocephalus. Aqueductal stenosis is not a
constant feature [50–53]. Hence, it has been proposed that
individuals carrying L1CAM mutations may have communicating hydrocephalus and that observed reductions in the
caliber of the aqueduct of Sylvius arise secondarily due to
compression from the enlarged ventricles. Intrafamilial
Neurogenetics
variation has also been observed with respect to L1CAMrelated hydrocephalus. Differences in the occurrence and
severity of hydrocephalus have been documented among
related males carrying the same L1CAM mutation. For
example, within the same family, some males with an
L1CAM mutation may not exhibit hydrocephalus, while others
display moderate or severe hydrocephalus [49–51, 54–57].
L1cam knock-out (L1KO) mice bred on a 129/Sv
background do not exhibit gross hydrocephalus upon
standard pathological examination [6, 7]. However, a
subtle, but significant, dilation of the lateral ventricles has
been detected with high-resolution magnetic resonance
imaging [58]. Interestingly, breeding L1KO mice onto a
C57BL/6J background enhances the L1KO phenotype and
results in severe hydrocephalus [6, 48, 59]. This suggests
that the degree of ventricular enlargement strongly depends
on genetic background, consistent with the intra- and
interfamilial variability of hydrocephalus severity in
humans. Hence, we hypothesized that ventricle size is
regulated by loci that (1) are polymorphic between the 129/
Sv and C57BL/6J mouse strains and (2) genetically interact
with L1cam, to exacerbate the L1cam mutant phenotype.
Furthermore, we hypothesized that L1cam modifier loci
could affect hydrocephalus susceptibility and/or severity.
Susceptibility modifiers would determine the presence or
absence of the phenotype. The same genotype at a
susceptibility modifier could exhibit wide variations in
severity. In contrast, severity modifiers would affect the
magnitude or spectrum of the phenotype. Thus, for a
severity modifier to have an effect, the mouse must already
be predisposed to hydrocephalus.
In this study, we attempt to uncover the genetic basis of
the strain-specific severe hydrocephalus phenotype of
L1cam mutants. To work toward the identification
of L1cam modifier genes that contribute to hydrocephalus,
we performed genome-wide linkage analyses on hydrocephalic F2 L1cam mutants derived from L1cam 129S2/
SvPasCrlf (129S2) and C57BL/6J mutant mice. Candidate
susceptibility loci were detected using chi-square tests to
identify markers that deviated from Mendelian segregation
in F2 L1cam mutants. In addition, QTL analyses of
hydrocephalic F2 L1cam mutants, as well as chi-square
tests comparing mutant mice with moderate versus severe
hydrocephalus, were conducted to identify candidate loci
that contribute to the severity of the condition.
Materials and methods
Mice and phenotypic analyses
All animal experiments and procedures described in this
manuscript were approved by the University of Miami’s
Institutional Animal Care and Use Committee. To generate
L1camtm1.1(L1-6D)Lem (L1-6D) mutant mice, the sixth
immunoglobulin-like domain of L1cam (exons 12–14) was
knocked out using a Cre/lox approach [60]. L1-6D heterozygous females (L1-6D/+) were then bred to wild-type 129S2/
SvPasCrlf (129S2) or C57BL/6J (B6) males for at least 12
generations to generate congenic (Cg) mice. A twogeneration, outcross–intercross breeding scheme was next
performed to generate the F2 mice utilized for linkage
analyses. Briefly, male 129S2.Cg-L1-6D/Y and female B6.
Cg-L1-6D/+ mice were bred to produce F1 L1-6D 129/B6
mice. The cross was set up in this manner because male and
female B6.Cg-L1-6D mutants are poor breeders. An intercross
was then set up between F1 L1-6D/Y males and heterozygous
or homozygous F1 L1-6D females to generate F2 progeny.
Mice were euthanized with carbon dioxide and evaluated
for hydrocephalus postmortem. Hemi- and homozygous
129S2.Cg-L1-6D (N=36: 32 L1-6D/Y males and 4 L1-6D/
L1-6D females) and B6.Cg-L1-6D/Y (N=10 males) mice
were collected over a wide age-span, ranging from weanlings
[postnatal day 21 (p21)] to adults (four B6.Cg-L1-6D and
eight 129S2.Cg-L1-6D mice over 6 months of age). F1 L1-6D
mutant mice (N=16: 14 L1-6D/Y males and two L1-6D/L16D females) were killed between p21 and 3 months of age.
All of the F2 mice included in this study were killed p21-p39:
wild type (N=27: 25 males and two females), L1-6D/+ (N=38
females), and L1-6D mutants (N=1,058: 540 L1-6D/Y males
and 518 L1-6D/L1-6D females). In particular, 186 of these F2
L1-6D mutant mice (105 L1-6D/Y males and 81 L1-6D/L16D females) were killed during a tight window, p21–p30, and
utilized for linkage analyses. Once their phenotype was
evaluated (as described below), they were divided into two
categories: (1) hydrocephalic/affected (N=156: 93 males and
63 females) and (2) non-hydrocephalic/unaffected (N=30: 12
males and 18 females). Hydrocephalic F2 L1-6D mice were
further separated into two subgroups: (1) mice that participated in both the genome-wide scan and extension study (N=92:
54 males and 38 females) and (2) mice that solely participated
in the extension study (N=64: 39 males and 25 females).
Whole brains were dissected from animals, rinsed with
1× phosphate buffered saline, and fixed in 4% paraformaldehyde. At least 24-h postfixation, thick coronal sections
were prepared. Brains were positioned with the cortical
surface parallel to the horizontal plane. A razor blade,
placed rostral to the cerebral cortex, was used to secure
each brain in place. Meanwhile, approximately four 1-mm
sections were cut, perpendicular to the brain surface, along
the rostral-caudal axis of the cortex using a scalpel (BardParker no. 371610; Fig. 1a). Sections were then examined
under a dissecting microscope (Olympus SZ-40, ×1.5 and
Wild Heerbrugg MG50, ×16). Sections were rarely
damaged during dissection. Nevertheless, marred samples
were not included in the study.
Neurogenetics
Fig. 1 The incidence and severity of hydrocephalus varies
among L1-6D mutant mice. a
Diagram of a mouse brain (dorsal
view) and a traced coronal section. Dotted lines represent the
approximate positions of coronal
sections. ImageJ software was
used to trace the perimeters and
calculate the areas of the cerebral
cortex section and both lateral
ventricles. b A recently weaned
129.Cg-L1cam-6D mouse with
ventricles of normal size. c, d
Two different B6.Cg-L1-6D mutant mice with hydrocephalus: (c)
adult and (d) weanling. e–g
Three recently weaned F2 L1-6D
mutant mice demonstrating the
varying severity of hydrocephalus observed among affected
mutants in the F2 generation
Mice were initially classified as hydrocephalic (affected) or non-hydrocephalic (unaffected) based on the
appearance of the lateral ventricles. Mice were categorized as unaffected, if both of their lateral ventricles were
shaped like thin slits. However, if at least one lateral
ventricle was expanded, the mouse was categorized as
hydrocephalic. Microscope-attached digital cameras
(Nikon and RT Slider Diagnostic Instruments, Inc.) were
used to image one or two sections showing the lateral
ventricles.
Neurogenetics
The hydrocephalic phenotype of F2 L1-6D mice utilized
for linkage analyses was quantified as follows. For each
mouse, we chose the imaged section that showed the
greatest expansion of the lateral ventricles. The positions of
these sections corresponded to reference sections positioned
within +1.0 mm from Bregma [61]. Next, the perimeter of
the coronal section and its respective lateral ventricles were
manually traced with ImageJ software [62]. ImageJ was
then further utilized to calculate the areas of the traced
regions. This generated three measurements: coronal
section area, left ventricle area, and right ventricle area.
Left and right ventricle areas were added together to
calculate total lateral ventricular area. To account for
differences in brain size and standardize measurements, a
ratio was calculated for each brain: total lateral ventricular
area divided by coronal section area. Ratios were then
transformed, using a natural log scale [ln (1,000×ratio)], to
approximate a normal distribution for statistical analyses.
We refer to these transformed ratios as hydrocephalus
severity scores. Student’s t tests were performed to
determine whether there were significant differences in
hydrocephalus severity between the different cohorts, as
well as between the two sexes.
Genotyping
Genomic DNA was obtained from mouse toes and/or tails.
Tissue samples were digested overnight with Proteinase K. L16D genotyping was performed using PCR amplification and
the following primer set: forward 5′- CCAGCCAGGATCC
TAACAAAAGAC, reverse wild-type allele 5′- AGT
GATGCTGGCCTGCAAAG, and reverse knock-in allele
5′- AACCACACTGCTCGACCTG. In preparation for single
nucleotide polymorphism (SNP) genotyping, genomic DNA
was treated with a RNAse A/RNAse T cocktail (Ambion)
and purified using either Puregene Blood Core Kit B (Gentra
Systems, Qiagen) or DNeasy Blood and Tissue Kit (Qiagen,
Cat. no. 69504). DNA concentrations were measured with a
Nanodrop ND-1000 Spectrophotometer.
The genome-wide scan was performed at the Affymetrix
GeneChip laboratory of the Yale Microarray Center for
Research on the Nervous System, through the NIH
Neuroscience Microarray Consortium. Illumina’s Mouse
Low Density Linkage panel and Golden Gate Assay were
used to genotype 92 F2 L1-6D homo-/hemizygote pups
with hydrocephalus at 377 SNPs; 99% of the genotypes
were present. Genotypes were classified as 129S2 or
C57BL/6J, and SNP performance was evaluated by the
Mouse SNP Genotyping Service at Harvard University. The
genotypes of our 129S2 mice matched that of 129S1 mice
at all 375 working SNPs except one (rs13481099), which
appeared monomorphic between our 129S2 and C57BL/6J
mice. Ultimately, 246 informative SNPs were polymorphic
between 129S2 and C57BL/6J mice; the average genomic
distance between these SNPs was 9.9 Mb.
We verified that all F2 L1-6D mice carried the L1-6D
associated 129/Sv congenic interval by sequencing SNP
rs3157210, located between exons 1 and 2 in the L1cam
gene. The following primers were used to amplify a 495
base pair fragment containing rs3157210: forward
5′- CACTCCACAATAGCCACACAC and reverse 5′GCCCTCACTTCTTCTGTAAC. PCR products were
then treated with ExoSAP-IT (USB) to remove excess
primers and nucleotides, and they were sequenced with
Finnzymes’ Phusion High Fidelity PCR Master Mix with
HF buffer (New England Biolabs) and primer 5′GCTGTTGAGTCAAGACCTGG.
Seven candidate modifier regions were selected for
further analyses based on their linkage data. SNPs in these
regions were chosen using the information obtained from
the Mouse Phenome Database (http://phenome.jax.org).
Genotyping was performed by the High Throughput
Sequenom and Illumina Genotyping Facility at the Harvard
Partners Center for Genetics and Genomics utilizing iPLEX
Gold chemistry from Sequenom. Ultimately, 43 polymorphic SNPs, three to eight SNPs per candidate region,
passed the performance criteria set by the Harvard Partners
Center (Supplemental Table 3). These SNPs were employed
to genotype the 92 F2 L1-6D mice with hydrocephalus
originally utilized for the genome-wide scan, the additional
F2 L1-6D mice used for the extension study (64 F2 L1-6D
mice with hydrocephalus and 30 F2 L1-6D mice without
hydrocephalus), and mice from the two parental strains
(two B6.Cg-L1-6D mice and two 129S2.Cg-L1-6D mice).
Linkage analysis
Linkage analysis was performed using a total of 288 SNPs
(272 autosomal and 16 X-linked). Mice with hydrocephalus
and those without were analyzed separately. In addition, sex
was used as a covariate. Data from male and female mice
were analyzed separately, as well as combined and analyzed
together. Progeny 6.9.04 software was used to visualize
haplotypes.
To test for linkage to hydrocephalus susceptibility
modifier loci, genotype frequencies at each marker were
tested for departures from Mendelian segregation with chisquare goodness-of-fit tests, two degrees of freedom, using
SAS software (SAS Institute, Inc., Version 9.1.3) [63, 64].
Fisher’s exact tests were used to generate p values for each
marker. However, when the frequency of a genotype at a
marker is 0, an exact test cannot be performed. In those
cases (see Supplemental Table 4), an asymptotic chi-square
test was done.
Bonferroni adjustments (α=0.05 and α=0.005) were
used to correct p values for multiple testing across SNPs: α
Neurogenetics
divided by the number of markers. During the original
genome-wide scan, 229 autosomal markers were tested.
Hence, we used Bonferroni thresholds of α/229 to interpret
the data. During the extension study, 43 additional
autosomal SNPs were analyzed. To maintain stringent
standards, extension study data were corrected utilizing all
272 autosomal markers (43 extension study markers + 229
autosomal markers): α/272.
To interpret X-chromosome marker data, we had to
analyze the two markers positioned inside the L1-6D 129
associated region (gnfX.035.350 and rs13483765) separately from the rest (N=14), which are positioned outside of the
region. This distinction was necessary because F1 L1-6D/
L1-6D homozygote females were 129 homozygotes within
the region and F1 L1-6D/+ females were 129/B6. When
performing chi-square tests, we also separately analyzed
male and female F2 progeny derived from the two F1
crosses: (1) L1-6D/L1-6D homozygote females bred to L16D/Y males and (2) L1-6D/+ females bred to L1-6D/Y
males. Depending on the type of F1 cross, the genotype
frequencies expected for X-chromosome SNPs far away
from L1cam (unlinked) vary. F2 L1-6D hemi-/homozygotes
derived from F1 L1-6D/L1-6D females are expected to
exhibit the following genotype frequencies: females—50%
B6 homozygotes, 50% B6/129 and males—50% B6 hemizygotes, 50% 129 hemizygotes. On the other hand, 100%
of the F2 L1-6D/L1-6D females derived from F1 L1-6D/+
females are expected to be 129/B6, and 100% of the
generated F2 L1-6D/Y males are expected to be 129
hemizygotes.
R/qtl 1.07-12 software was used for QTL analyses to test
for linkage to severity modifier loci [65]. Single-locus and
two-QTL, two-dimensional genome-wide scans, using a
multiple imputation method with 2.5-Mb steps over the
entire genome and 200 imputations, were performed
utilizing 229 autosomal genome-wide markers. In addition,
permutation tests were applied to single-locus and two-QTL
scans as recommended for an intercross (http://www.rqtl.
org/tutorials/). Multiple QTL mapping functions were also
implemented (source: http://www.rqtl.org/multqtlfunc.R).
Using recommended LOD thresholds [65], the results from
the two-dimensional scan were summarized. The twodimensional scan summary was then used to fit a multiple
QTL model utilizing the “fitqtl” function in R/qtl.
As described in Broman et al. [66], R/qtl was also used
to perform genome-wide scans that analyzed 14 Xchromosome markers along with all 229 autosomal
markers. The X-chromosome markers gnfX.035.350 and
rs13483765, which are positioned within the 129 congenic
interval present in all L1-6D mutants, were excluded.
Thirty-nine F2 L1-6D hemi-/homozygous mutants derived
from F1 L1-6D/L1-6D females were used for the
X-chromosome R/qtl study. F2 L1-6D mutants derived
from F1 L1-6D/+ females were excluded from the analysis
because their siblings were not genotyped and the missing
data would confound analysis.
In addition, R/qtl was used to analyze data from the
extension study. Three sets of analyses were conducted.
First, the cohort of 64 F2 L1-6D mutants with hydrocephalus (utilized solely for the extension study) was examined
at the 43 extension study SNPs. Second, the complete
group of 156 F2 L1-6D hydrocephalic mutant mice (64 +
92 from the original genome scan) was examined at the 43
extension study SNPs. Lastly, the original 92 F2 L1-6D
mice with hydrocephalus utilized for the genome-wide scan
were analyzed with the complete set of 272 autosomal
markers (229 original SNPs + 43 new validation study
SNPs). In addition, cohorts were divided by sex, and males
and females were analyzed separately. Single-locus effect
plots and two-loci interaction effect plots were generated
with R/qtl.
As an alternative test for severity modifiers, F2 L1-6D
hydrocephalic mutant mice were classified as having
moderate or severe hydrocephalus based on whether their
hydrocephalus severity score was higher or lower than the
median hydrocephalus severity score of all F2 L1-6D
hydrocephalic mutant mice. A chi-square test with two
degrees of freedom was performed to compare genotype
frequencies between the moderate and severe hydrocephalus groups.
Microarray analysis
Gene Expression Omnibus (GEO) data set GDS 1490 was
downloaded from http://www.ncbi.nlm.nih.gov/geo. Normalized TeraGenomics-calculated signal intensities and
detection calls were obtained for 91 Affymetrix Murine
Genome U74AV2 gene chips comprising 24 neural tissues
in C57BL/6J and 129S6/SvEvTac mice [67]. Replicates
were pooled, and mean fold changes were calculated using
Spotfire DecisionSite (TIBCO, Version 9.1.1). Statistical
analyses were not performed to determine significant
differences because of the small sample size. Each
condition had only two replicates, except for the CA1
region of the hippocampus, which had three replicates.
In addition, our lab performed a microarray experiment:
GEO data set GSE13984. RNA was purified from the
cerebellums of three postnatal day11 congenic 129S2
L1cam knock outs (129S2.Cg-L1camtm1Sor) and three of
their wild-type siblings utilizing Qiagen’s RNeasy Mini kit
(no. 74104). Therefore, each genotype, L1KO and wild
type, had three replicates. Samples were processed at the
Case Western Reserve University Cancer Center Gene
Expression Array Core Facility and hybridized to Affymetrix Murine Genome U74AV2 gene chips. Affymetrix
GeneChip Operating Software was used to process gene
Neurogenetics
chip images and generate MAS 5.0 signal intensities and
detection calls. Replicates were normalized by the mean
signal value of all probes and then pooled using Spotfire
DecisionSite. Significance was determined by performing
an unpaired two-tailed Student’s t test on normalized signal
intensities.
Results
Strain-specific modifiers contribute to hydrocephalus
in L1-6D mice
The incidence of hydrocephalus differs among strains of
L1cam mutant mice. Severe hydrocephalus is evident in
C57BL/6J, but not 129/Sv, L1cam mutants. To identify
modifier loci that interact with L1cam to cause X-linked
hydrocephalus, we performed linkage analyses on L1cam
mutants derived from the breeding of 129S2 and C57BL/6J
mice. To generate sufficient mice for the study, we utilized
L1-6D knock-in mice rather than L1KOs. Unlike L1KOs,
which fail to express L1, L1-6D knock-in mice express a
mutant protein in which the sixth Ig-like domain of L1 has
been deleted [60]. This Ig-like domain, which contains two
Arg-Gly-Asp (RGD) sequences, is required for L1 homophilic (L1-L1) and L1-integrin binding in vitro; yet, despite
the deletion, the L1-6D protein retains some function. Its
interactions with Neurocan and Neuropilin-1 are preserved,
and in contrast to L1KOs, L1-6D mutants develop a normal
corticospinal tract. Nevertheless, L1-6D mutants, like
L1KOs, have been observed to exhibit hydrocephalus, on
the C57BL/6J, but not 129S2 background [60].
Consistent with these findings, we found that the incidence
of hydrocephalus differs between our congenic L1-6D
C57BL/6J (B6.Cg-L1-6D) and 129S2 (129S2.Cg-L1-6D)
mice. Coronal sections were prepared from L1-6D mutant
(L1-6D/Y hemizygotes and L1-6D/L1-6D homozygotes)
mice of varying ages, p21-adults (Fig. 1a). Dramatic
hydrocephalus was evident in all B6.Cg-L1-6D mutant mice
analyzed (N=10; Fig. 1c, d). However, hydrocephalus was
seldom seen in 129S2.Cg-L1-6D mutant mice (Fig. 1b).
Only two of 36 mice (5.6%) exhibited mild hydrocephalus
(data not shown). Hence, our congenic L1-6D lines could
validly be used to identify modifier loci that (1) are
polymorphic between the two strains and (2) influence the
manifestation of L1cam X-linked hydrocephalus.
In addition, the use of L1-6D, rather than L1KO mice,
was advantageous because L1KO (L1camtm1Sor) males are
infertile (Supplemental Text 1). On the other hand, 129S2.
Cg-L1-6D hemizygous males (L1-6D/Y) are fertile, and
their fertility permits the generation of L1-6D homozygous
females, which are also fertile. The ability to breed L1-6D
hemi- and homozygous mutants allowed us to generate
sufficient mutants for linkage mapping, despite the significant pre- and postnatal mortality caused by L1cam
mutations (Supplemental Text 2).
This early mortality phenotype, prevalent among both
L1-6D 129S2 and C57BL/6J mice, contrasts with the severe
hydrocephalus phenotype, which is strain specific. Hence,
while modifier loci may influence both mortality and
hydrocephalus in L1cam mutants, the data suggest that
modifier loci that contribute to non-lethal severe hydrocephalus play a major role in the development of
hydrocephalus. Since we evaluated all mice for hydrocephalus after weaning, our study was designed to identify this
latter class of modifiers.
As illustrated in our breeding scheme (Fig. 2), we first
mated B6.Cg-L1-6D/+ females to 129S2.Cg-L1-6D/Y
males. We evaluated 16 F1 L1-6D hemi-/homozygous
mutants, between p21 and 3 months of age, for the
phenotype. None exhibited hydrocephalus, demonstrating
that primary (main-effect) hydrocephalus-causing modifiers
do not act in a dominant manner. This observation
additionally suggests that if a main-effect modifier is
inherited from the C57BL/6J strain, it is not sex-linked. If
it were, we would expect F1 L1-6D/Y males to exhibit
hydrocephalus since their X chromosome, which carries the
L1-6D mutation, was inherited from the C57BL/6J strain.
Next, we set up an intercross between F1 L1-6D/Y 129/B6
males and F1 L1-6D homo- or heterozygote 129/B6 females.
We implemented this breeding scheme, which provided
opportunities for meiotic recombination between the 129S2
and C57BL/6J chromosomes of both parents, to facilitate
linkage analysis. Based on this cross (Fig. 2), if the C57BL/
6J allele of one recessive, fully penetrant modifier gene
caused hydrocephalus, one would expect 25% of F2 L1-6D
mutants to be B6 homozygotes at the modifier locus and
consequently exhibit hydrocephalus. We evaluated 1,058 F2
L1-6D mutants, between p21 and p39, for the phenotype.
Strikingly, 31.9% of them exhibited hydrocephalus with
variable severity (Fig. 1e–g): N=338 hydrocephalic F2 L16D mutants (199 L1-6D/Y males and 139 L1-6D/L1-6D
females). In contrast, the phenotype was not observed in any
F2 wild-type siblings (N=27). While 5.3% of F2 L1-6D/+
Fig. 2 Breeding scheme used to map modifier loci that contribute to
L1cam X-linked hydrocephalus
Neurogenetics
heterozygote siblings (two of 38) exhibited hydrocephalus,
the phenotype was much less severe than that exhibited by
F2 L1-6D hemi-/homozygotes (data not shown).
Based on these data, we hypothesized that main-effect
hydrocephalus susceptibility modifiers, which determine
the presence or absence of the hydrocephalus phenotype in
L1cam mutants, act in a recessive manner. In addition, since
129S2 L1-6D mutants seldom exhibit hydrocephalus, we
expected C57BL/6J alleles at main-effect loci to be
primarily responsible for causing the phenotype. Furthermore, since the severity of the phenotype varied among F2
mutants, we hypothesized that hydrocephalus is a complex
trait, whose severity is affected by multiple, interacting loci.
Genome-wide scan identifies multiple candidate modifier
loci
To identify the chromosomal location of L1cam-associated
hydrocephalus modifier loci, we conducted a linkage study
with an initial genome-wide scan on 92 F2 L1-6D mice with
hydrocephalus. This group consisted of 55 L1-6D/Y males
and 37 L1-6D/L1-6D females, whose ages ranged from
postnatal day21 to 30. To quantify the phenotype, a
representative coronal section was analyzed for each
individual. The areas of both lateral ventricles and the brain
section were measured. Total ventricular area was then
divided by total brain area to normalize measurements, and
values were transformed for statistical analyses. We refer to
these transformed values as the hydrocephalus severity score.
Figure 3a illustrates the distribution of hydrocephalus
severity among the 92 F2 L1-6D mutants. Mice were also
separated by sex to compare the distribution of hydrocephalus severity scores in males versus females (Fig. 3b–c). In
this cohort of 92 mice, males tended to exhibit slightly more
severe hydrocephalus than females, p<0.05.
Each mouse was genotyped at 245 informative SNPs,
which were polymorphic between the C57BL/6J and 129S2
strains. Since our goal was to uncover both hydrocephalus
susceptibility and severity modifiers, we utilized two
distinct approaches for data analyses. Chi-square tests for
Mendelian segregation were used to identify candidate
susceptibility loci. Meanwhile, QTL analysis was employed
to detect candidate severity loci.
We commenced our search for candidate susceptibility
markers by testing each marker for linkage to the hydrocephalus phenotype. Based on the breeding scheme (Fig. 2), we
hypothesized that autosomal markers not linked to hydrocephalus would exhibit the expected 1:2:1 Mendelian ratio of
B6/B6:B6/129:129/129. In contrast, we expected the
genotype frequencies of markers linked to hydrocephalus to
deviate from Mendelian segregation. Since hydrocephalus
was universally seen in congenic C57BL/6J L1-6D mutants,
but only seldom seen in congenic 129S2 L1-6D mutants, we
hypothesized that the frequency of B6 homozygotes at
autosomal SNPs linked to hydrocephalus would significantly
exceed 25%. A chi-square goodness-of-fit test was utilized to
compare the observed genotype frequencies at each SNP
(N=229 autosomal SNPs) with those expected based on
Mendelian segregation.
Four genomic regions, located on chromosomes 2, 3, 4
and 5, exhibited deviations from Mendelian segregation,
p<0.001 (Table 1). In these regions, a greater percentage of
F2 L1-6D mutants were B6 homozygotes than expected.
Rather than 25% of the mice being B6 homozygotes, more
than 42% were B6 homozygotes. This is consistent with the
hypothesis that B6 alleles at modifier loci near these SNPs
contribute to hydrocephalus. Based on the stringent criteria
proposed by Lander and Kruglyak for genome-wide
significance using F2 intercross progeny (two degrees of
freedom), these findings would be classified as evidence
of suggestive linkage because p < 1:6 " 10#3 [68]. Moreover, multiple markers on chromosomes 2, 3 and 5 surpassed
Lander and Kruglyak’s significant linkage threshold of p <
5:2 " 10#5 and retained significance after a Bonferroni
adjustment (α=0.005) was used to correct for multiple
testing of the null hypothesis.
Chi-square tests for deviations from Mendelian segregation were also separately performed on males and females.
The same regions on chromosomes 2, 3, and 5, which were
specified above, exhibited deviations from Mendelian
segregation in both sexes, p<0.01 (Supplemental Table 1).
In contrast, linkage evidence for the region on chromosome
4 differed between the sexes. At one marker (rs3023025),
females showed no evidence of linkage
! (p>0.05), while
"
males exhibited suggestive linkage p ¼ 9:25 " 10#4 .
However, this p value did not pass Bonferroni correction
(α=0.05). Hence, our data do not provide evidence that loci
on chromosome 2, 3, 4, or 5 contribute to the development
of hydrocephalus in one sex but not the other.
We then performed QTL analysis to map candidate
hydrocephalus severity loci. R/qtl software was used to
analyze the data collected from the aforementioned 92 F2
L1-6D hydrocephalic mutants. Associations between marker genotypes at 229 genome-wide autosomal SNPs and
hydrocephalus severity scores were investigated. QTL
analysis showed no main-effect loci linked to hydrocephalus severity. We found a linkage peak on chromosome 10
at 119 Mb (mCV25429984) with a LOD score of 2.9
(Fig. 4). However, a permutation test of the single-locus
scan did not indicate significant evidence of linkage at this
locus. A two-loci scan was then performed. Three pairs of
loci had overall LOD scores above 6 (full LOD) and
interaction LOD scores, which measure the statistical
interaction between loci, above 3.5: (1) chromosome 4 at
153.7 Mb and chromosome 9 at 112.2 Mb, (2) chromosome
5 at 109.2 Mb and chromosome 9 at 102.2 Mb, and (3)
Neurogenetics
Fig. 3 Histograms illustrating the distribution of hydrocephalus
severity among the different cohorts of F2 L1-6D mutant mice
analyzed during the study. Hydrocephalus severity is plotted against
the number of mice with scores in the specified range. Hydrocephalus
severity increases along the x-axis. a–c Severity histograms for the 92
F2 L1-6D mutants analyzed in the genome-wide scan. a All mice. b, c
Group segregated by sex: b males and c females. d Severity histogram
for the 64 additional F2 L1-6D mutant mice used in the extension
study. e Severity histogram representing the entire set (92+64) of 156
F2 L1-6D mutant mice with hydrocephalus
chromosome 7 at 101.6 Mb and chromosome 9 at 87.2 Mb
(Supplemental Table 2). However, again, the permutation
test revealed no significant results for this two-QTL scan.
Furthermore, no QTLs were detected in males nor females
when data were segregated by sex.
crosses, (1) L1-6D/L1-6D homozygote females bred to L16D/Y hemizygote males and (2) L1-6D/+ females bred to
L1-6D/Y hemizygote males, were analyzed separately. As
expected, rs13483765, a SNP in the L1-6D associated 129
region positioned 16.3 Mb away from L1cam, showed
linkage to hydrocephalus. Genotype frequencies at
rs13483765 deviated from the Mendelian segregation
expected of an unlinked marker (a marker far away from
L1cam). A greater percentage of F2 L1-6D/Y males
originating from F1 L1-6D/L1-6D females were 129 hemizygotes than expected for an unlinked X-chromosome
marker; 129 hemizygosity was observed in 87% of the mice
rather than in the expected 50%, p ¼ 4:88 " 10#4 ðN ¼ 23Þ.
A valid analysis of rs13483765 in F2 L1-6D/L1-6D females
derived from either cross, as well as in F2 males derived
from F1 L1-6D/+ females, was not possible. The sample
sizes in these subgroups were too small.
A genome-wide QTL scan was also performed, including
X-chromosome markers (14 X-chromosome markers + 229
X-chromosome analyses
Analyses of the X chromosome required us to take into
consideration that 129/Sv DNA was originally utilized to
generate the L1-6D knock-in allele [60]. Consequently, all
L1-6D C57BL/6J mice carry a 129 congenic interval, which
includes the L1cam locus (positioned at 71.1 Mb). Sequencing revealed that this L1-6D associated 129 region is
<25 Mb, and it lies between SNPs rs24884396 and
rs3166693 (genome coordinates, 46.6–71.5 Mb).
Chi-square goodness-of-fit tests were performed on Xchromosome markers to test for departures from Mendelian
segregation. Male and female progeny from the two F1
Neurogenetics
Table 1 Candidate modifier
loci identified through genomewide chi-square tests for Mendelian segregation
N=92 F2 L1-6D mice with hydrocephalus. The values reported
in the Genotype columns denote
the number of mice with the
specified genotype
a
Exact p value passed Bonferroni
correction with α=0.05 and 229
SNPs
b
Exact p value passed
Bonferroni correction with
α=0.005 and 229 SNPs
SNP
Chr
Mb
Genotype
Exact p value
B6/B6
B6/129
129/129
rs4223428
rs8251635
rs3655895
rs13476832
rs6376291
rs13476892
rs3712766
rs3691114
rs6239288
rs3683507
rs13477223
rs3688780
2
2
2
2
2
2
2
3
3
3
3
3
120.322
125.128
144.359
149.561
153.967
165.222
165.836
59.630
60.315
79.005
80.666
86.177
39
41
41
42
42
42
42
43
44
45
45
42
43
41
40
40
40
36
34
31
31
30
29
32
10
10
11
10
10
14
15
17
17
17
18
18
1.07E−04a
1.98E−05b
3.10E−05a
8.53E−06b
8.53E−06b
2.60E−05a
2.31E−05a
9.06E−06b
4.50E−06b
1.36E−06b
1.13E−06b
3.52E−05a
rs4138858
rs3023025
rs13478069
rs3726547
rs3663141
rs3662161
gnf05.120.578
3
4
4
5
5
5
5
87.446
142.772
154.877
108.813
117.205
117.909
124.103
41
39
38
41
41
42
42
33
39
37
38
41
40
43
18
14
16
13
10
10
7
9.73E−05a
3.97E−04
9.87E−04
5.36E−05a
8.53E−06b
1.98E−05b
2.14E−06b
autosomal markers). The L1-6D associated 129 region was
excluded because the 129/B6 genotype of parental females
complicated the analysis. Furthermore, only 39 F2 L1-6D
hemi-/homozygous mutants derived from F1 L1-6D/L1-6D
females were used for the X-chromosome QTL study. F2 L16D mutants derived from F1 L1-6D/+ females were not used
because their siblings, non-hydrocephalic wild-type males and
L1-6D/+ females, were not genotyped with the genome-wide
linkage panel. Since these F2 wild-type males and L1-6D/+
females would have inherited their X chromosomes from the
B6 parental strain, the missing data would confound analysis.
Single- and two-loci scans did not identify any Xchromosome markers linked to hydrocephalus severity.
This result was not surprising, as the sample size was
small. Furthermore, since the examined mice were all L16D hemi-/homozygotes with hydrocephalus of varying
severity, there was no genetic variation at the locus to
correlate with phenotype severity.
Fig. 4 LOD plot of a genome-wide scan for main-effect loci linked to hydrocephalus severity; 92 hydrocephalic F2 L1-6D mutants were
genotyped at 229 informative autosomal markers
Neurogenetics
cohort, a lower percentage of mice in the new group had
hydrocephalus severity scores less than 2 (Fig. 3d). While
13% of the mice in the original cohort scored below 2, only
1.6% of the mice in the new cohort scored below 2. We
preferentially chose mice with the most severe phenotype
for the extension study because we thought that they might
be enriched for individual alleles or combinations of alleles
that have strong effects.
Extension study design
An extension study was next conducted to determine
whether stronger linkage evidence could be obtained for
the candidate modifier loci identified through the genomewide scan. Seven genomic regions were analyzed in total.
Two of these regions, located on chromosomes 4 and 5,
were originally identified by both our chi-square analysis
and two-loci QTL scan. Candidate modifier regions on
chromosomes 2 and 3 were selected for further study based
on their chi-square test results, while those on chromosomes 7 and 9 were chosen based on their two-loci QTL
scan data. Lastly, a region on chromosome 10, surrounding
rs3654717 (106.9 Mb), was further studied for two reasons.
First, it was detected by the two-loci QTL scan, albeit with
low LOD scores. When paired with a locus on chromosome
7 (60.2 Mb), the chromosome ten locus at 106.9 Mb had a
full LOD score of 5.45 and an interacting LOD score of
0.261. Second, it was near the highest peak of the singlelocus QTL scan.
For the extension study, 43 informative SNPs
(Supplemental Table 3), dispersed throughout the candidate
modifier regions, were used to genotype two new groups of
mice: (1) a control set of 30 F2 L1-6D hemi/homozygous
mutant siblings without hydrocephalus (unaffecteds) and
(2) 64 additional F2 L1-6D hemi-/homozygous mutant mice
with hydrocephalus (affecteds). Within this new cohort, the
severity of hydrocephalus was not significantly different
between males and females (p> 0.05). Moreover, the
severity of hydrocephalus exhibited by the new group of
64 F2 L1-6D mutants was not significantly different from
that displayed by the original cohort of 92 F2 L1-6D
mutants (p>0.05). Nevertheless, compared to the original
Identification of a L1cam hydrocephalus modifier locus
on chromosome 5
The extension study replicated the results obtained for the
candidate modifier region on chromosome 5. A chi-square
test for deviation from Mendelian segregation was performed on the new cohort of 64 F2 L1-6D mice with
hydrocephalus utilizing the chromosome 5 SNPs listed in
Supplemental Table 3. All eight new markers exhibited
deviations from Mendelian segregation; B6 homozygotes
were more prevalent than other genotypes (Table 2). Six
adjacent markers (100.1–120 Mb) demonstrated evidence
for significant linkage, surpassing the p < 5:2 " 10#5
threshold proposed by Lander and Kruglyak [68]. Moreover, five of these markers retained significance following
Bonferroni correction (α=0.005).
The group was also divided by sex, and the region was
separately tested for linkage among males and females.
Original genome-wide scan results were replicated.
Markers in the chromosome 5 candidate modifier region
deviated from Mendelian segregation in both sexes,
p<0.005 (Supplemental Table 4).
Moreover, combined analysis of all 156 F2 L1-6D
mutants with hydrocephalus (64 from the extension study
Table 2 Extension study chi-square tests for Mendelian segregation on chromosome 5
Marker
Position (Mb)
64 new F2 L1-6D mutants with
hydrocephalus
All 156 F2 L1-6D mutants with
hydrocephalus
30 F2 L1-6D mutants without
hydrocephalus
Genotype
Exact
Genotype
Exact
Genotype
p value
B6
p value
B6
H
B6
H
129
H
129
Exact p value
129
rs3688859
rs3694887
rs3654076
rs3665124
rs3657810
rs3653889
rs3710934
100.055
104.164
106.519
110.024
113.832
119.958
121.681
31
34
33
32
31
30
29
29
27
27
29
30
29
30
3
3
3
3
3
4
5
6.12E−06a
2.92E−07a
6.64E−07a
3.17E−06a
6.18E−06a
2.58E−05b
1.29E−04b
70
75
73
73
72
69
70
72
68
69
67
71
72
73
13
13
13
16
13
13
13
2.01E−09a
4.04E−11a
1.41E−10a
6.64E−10a
4.23E−10a
3.30E−09a
1.97E−09a
6
7
7
5
5
6
6
17
16
16
18
19
19
18
7
7
7
7
6
5
6
0.791
0.943
0.943
0.492
0.407
0.407
0.591
rs6345336
129.158
24
34
5
0.0025
65
77
13
5.06E−08a
6
18
6
0.591
Genotype: B6 = B6/B6, H = B6/129, 129=129/129
a
Exact p value passed Bonferroni correction with α=0.005 and 272 SNPs
b
Exact p value passed Bonferroni correction with α=0.05 and 272 SNPs
Neurogenetics
Table 3 LOD scores for interacting loci pairs
Interacting pair
Chr5 at 106.7 Mb : Chr9 at
102.2 Mb
Chr7 at 101.6 Mb : Chr9at
89.7 Mb
Full LOD
score
Interaction LOD
score
7.23
4.58
6.62
3.70
Data from genome-wide two-loci QTL scan performed on 92
hydrocephalic F2 L1-6D mice using 272 autosomal markers
and 92 from the original genome-wide scan) resulted in the
significant linkage of the eight extension study SNPs. All
markers passed Bonferroni correction (α=0.005), and the
lowest p value was exhibited by rs3694887 (104.2 Mb),
p ¼ 4:04 " 10#11 . Significant linkage was also observed in
both sexes (Supplemental Table 4). Taken together, the
genome-wide scan and extension study results confirm
linkage of this region to hydrocephalus. We denote our
newly discovered modifier locus on chromosome 5, which
genetically interacts with L1cam to contribute to hydrocephalus as L1cam hydrocephalus modifier 1 (L1hydro1).
Based on our chi-square results and haplotype visualization,
we propose that L1hydro1 is located in the region between
genome coordinates 100–129.2 Mb.
Interestingly, no single genotype in the L1hydro1
region was shared between all hydrocephalic mice. This
suggests that the L1hydro1 modifier does not, on its own,
account for all hydrocephalus cases. In addition, the
genotype frequencies of F2 L1-6D mutant mice without
hydrocephalus did not deviate from Mendelian segregation at L1hydro1 (Table 2). If harboring a B6 homozygote
genotype at L1hydro1 causes hydrocephalus in L1cam
mutants, few, if any, unaffected L1-6D mutants should be
B6 homozygotes at L1hydro1. Rather, one would expect
more hetero- or 129 homozygotes at L1hydro1 in
unaffected L1-6D mice, yet this was not seen. Taken
together, these last two observations suggest that L1hydro1
exhibits incomplete penetrance, i.e., some individuals
carrying the allele that contributes to hydrocephalus fail
to exhibit the trait.
Extension study analyses of candidate modifiers
on chromosomes 2, 3, and 4
The extension study, performed on 64 new F2 L1-6D
mutants with hydrocephalus, failed to detect linkage to
markers on chromosomes 2 and 3. In contrast to data from
the original genome-wide study, genotypes with significant
deviations from Mendelian segregation were not observed
(p>0.05). Furthermore, although four markers within a
region on chromosome 4 (134.7–149.8 Mb) exhibited
p<0.05, significance was not retained after Bonferroni
correction (α=0.05). Hence, the extension study also did
not validate linkage of the chromosome 4 locus.
Nevertheless, when we analyzed genotype frequencies
for deviations from Mendelian segregation at the extension
study SNPs utilizing our entire set of 156 F2 L1-6D
hydrocephalic mutants, multiple markers on chromosomes
2, 3, and 4 surpassed Lander and Kruglyak’s proposed p <
1:6 " 10#3 threshold for suggestive linkage. Furthermore,
they retained significance after Bonferroni adjustment
(α=0.05). One marker on chromosome 2 (rs13476825)
and two markers on chromosome 4 (rs6191908 and
rs3675987) even exhibited evidence for significant linkage,
p < 5:2 " 10#5 (Supplemental Table 5).
Thus, while the extension study did not replicate results,
the loci on chromosomes 2, 3, and 4 continued to exhibit
evidence of linkage when the extension study set of mice
were analyzed together with the original set of 92 mice.
Hence, the regions remain candidate modifier loci. Indeed,
while the loci may be false positives, it is also possible that
these loci have smaller effects than the L1Hydro1 locus and
that our extension study lacked sufficient power to detect
linkage. Alternatively, it may be that the loci contribute to
milder forms of hydrocephalus and that the distribution of
hydrocephalus severity within the extension study group
prohibited detection.
QTL analyses of extension study data
Extension study data were also analyzed to determine
whether candidate modifier loci affect hydrocephalus
severity. QTL analyses were performed, using the 43 new
extension study SNPs, on two groups: (1) the extension
study cohort of 64 hydrocephalic F2 L1-6D mutant mice
and (2) the entire set of 156 F2 L1-6D hydrocephalic mice.
No linkage evidence was found. Individual loci did not
show linkage, nor did any pair of interacting loci meet the
recommended thresholds [65] for inclusion into a two-QTL
model. While it is possible that the interacting loci on
chromosomes 4, 5, 7, and 9, which were detected by the
original two-loci QTL genome-wide scan, were false
positives, the distribution of hydrocephalus severity in the
new cohort may have prevented replication. In addition, our
extension study may have been underpowered to identify
multiple QTL that contribute to hydrocephalus severity but
do not exhibit a strong individual effect.
We then reanalyzed our original 92 F2 L1-6D hydrocephalic mice with our entire set of 272 autosomal markers
(246 original genome scan markers + 43 new extension
study markers). Genome-wide single- and two-loci scans
were conducted. Individual loci did not demonstrate
linkage. In addition, this analysis revealed no evidence for
a two-QTL interaction involving the chromosome 4 locus.
Furthermore, while the three regions on chromosomes 5
Neurogenetics
(106.7 Mb), 7 (99.1 Mb), and 9 (92.2–102.2 Mb) again met
thresholds for inclusion into a two-QTL model (Table 3), a
permutation test indicated that the loci pairs still did not
exhibit significant evidence of linkage.
We then examined the simultaneous effects of multiple
QTLs by incorporating the four loci on chromosomes 5, 7,
and 9 into a multiple QTL model. Individual loci, as well as
pair-wise gene interactions, exhibited p values less than
0.005 (Table 4). Moreover, the full QTL model could
explain 48% of! the phenotypic" variance associated with
hydrocephalus p < 5:3 " 10#6 . Hence, we continued to
consider these three regions on chromosomes 5, 7, and 9 as
candidate severity QTL.
To better understand the relationship between genotypes
at these candidate QTL and hydrocephalus severity, we
generated single-locus effect plots, comparing genotype to
hydrocephalus severity (Fig. 5a–c). Mice with a B6/B6
genotype at marker rs3654076, located within the L1hydro1
region on chromosome 5, tended to have more severe
hydrocephalus than mice with the 129/B6 or 129/129
genotype (Fig. 5a). In contrast, mice with a 129/129
genotype at rs13479392, the marker linked to the candidate
chromosome 7 QTL, tended to exhibit more severe
hydrocephalus than mice with other genotypes at that
marker (Fig. 5b). Single-locus effect plots for the two
chromosome 9 markers, rs4138352 and rs3657346 (92.2–
102.2 Mb), did not reveal substantial differences in
hydrocephalus severity between different genotypes
(Fig. 5c). However, interaction effect plots, which depict
how genotypes at two-loci influence the phenotype, indicate
that the severity of hydrocephalus exhibited by mice carrying
a B6/B6 genotype at both chromosome 9 markers is
augmented by B6 homozygosity at chromosome 5
(rs3654076) or 129 homozygosity at chromosome 7
(rs13479392; Fig. 5d–e). Taken together, all these effect plots
suggest that C57BL/6J mice carry an allele on chromosome
5, within L1hydro1, that exacerbates the hydrocephalus
phenotype. Furthermore, genotypes at L1hydro1 may interact
with genotypes at the candidate chromosome 9 QTL to affect
severity. In turn, genotypes at the candidate chromosome 9
QTL may additionally interact with those at the third
candidate severity QTL on chromosome 7.
A candidate hydrocephalus severity modifier
on chromosome 10
As an additional approach to identify hydrocephalus severity
modifiers, we conducted chi-square tests, comparing groups
stratified by severity. Affected (hydrocephalic) mice were
divided into two groups based on severity. First, all 156
hydrocephalic F2 L1-6D mutant mice were used to calculate
the median hydrocephalus severity score, median = 3.04.
Those with a hydrocephalus severity score lower than the
median were classified as having moderate hydrocephalus,
while those with a hydrocephalus severity score higher than
the median were classified as exhibiting severe hydrocephalus. The null hypothesis was that Mendelian segregation
would be observed in both the moderate and severe
hydrocephalus groups. Next, all 156 mice were analyzed
at the 43 new markers (Supplemental Table 3). Interestingly, a locus on chromosome 10 deviated from Mendelian
segregation in the moderate (p<0.05), but not severe
hydrocephalus group (p<0.09). A chi-square test, comparing the genotype frequency (B6/B6:B6/129:129/129)
between the moderate and severe hydrocephalus groups,
revealed a significant difference, p<0.005. More mice with
moderate hydrocephalus were 129, rather than B6, homozygotes within the chromosome 10 candidate modifier
region, and vice versa, more mice with severe hydrocephalus were B6, rather than 129 homozygotes (Table 5).
These data suggest that a candidate severity modifier locus
is located on chromosome 10 (104.4–108.2 Mb) and that
moderate hydrocephalus correlates with 129 homozygosity
at the locus.
Curiously, when mice were segregated into four groups
according to sex and hydrocephalus severity (males or
Table 4 Multiple QTL model for hydrocephalus severity
Chromosomal location (Mb)
Nearest marker
Variance (%)
F value
p value
Chr5 at 106.7
Chr7 at 99.1
Chr9 at 92.2
rs3654076
rs13479392
rs4138352
17.88
15.58
16.06
2.33
3.76
3.89
0.00083
0.00244
0.00194
Chr9 at 102.2
Chr5 at 106.7:Chr9 at 102.2
Chr7 at 99.1:Chr9 at 92.2
% variance explained by full model
rs3657346
17.30
14.97
13.53
48.0
4.19
5.44
4.92
0.00109
0.00067
0.00142
5.3×10−6
N=92 F2 L1-6D mutants with hydrocephalus. Variance indicates the percentage of phenotypic variance explained by the specified locus,
interacting loci, or full model (including interaction effects). F values were generated by R/qtl through ANOVA analysis. p values reflect the
effects that removal of the specified locus or interacting loci have on the model
Neurogenetics
Fig. 5 a–e Effect plots depicting the effects of genotype on
hydrocephalus severity at candidate QTL. Hydrocephalus severity
increases along the y-axis. a–c Single-locus effect plots. Genotypes for
the SNP listed at the top of each graph are indicated along the x-axis.
The plot for the chromosome 9 SNP rs4138352 is similar to the one
for rs3657346 (data not shown). d, e Interaction effect plots depict
how the combinations of genotypes at two-loci influence severity.
Genotypes for the SNP at the bottom of each graph are indicated along
the x-axis. Genotypes for the second, interacting SNP (listed at the top
of the graph) are coded according to the legend. Error bars
represent ±SEM
females with moderate or severe hydrocephalus), only
females with moderate hydrocephalus had genotypes that
deviated from Mendelian segregation at the candidate
chromosome 10 modifier locus, p<0.05 (Table 5). In this
group, the incidence of 129 homozygotes was greater than
expected. In contrast, we found no evidence that genotypes
on chromosome 10 affect the phenotype of females with
severe hydrocephalus or males. Nevertheless, the data do
suggest that hydrocephalus severity differs between the
sexes. A Student’s t test, utilizing the entire set of 156
hydrocephalic F2 L1-6D mutants, indicated that females
exhibit a slightly milder phenotype than males (p<0.007).
Table 5 Chi-square tests for Mendelian segregation in F2 L1-6D mice with moderate versus severe hydrocephalus
Group
All mice
Females
Males
Phenotype
Moderate
Severe
Moderate
Severe
Moderate
Severe
Genotype at rs13480752
B6/B6 (%)
129/B6 (%)
129/129 (%)
14.3
29.1
8.1
34.6
20.5
26.4
50.7
57
51.4
53.9
48.7
58.5
35
13.9
40.5
11.5
30.8
15.1
Exact p value
p value: genotype frequency comparison
0.034
0.075
0.021
0.236
0.689
0.248
0.003
0.006
0.196
N=156 affected F2 L1-6D mice. Percentages represent mice carrying the designated genotype in each hydrocephalus severity category. The exact
p value was generated from a Fisher’s exact test for deviation from Mendelian segregation within each phenotypic class. The second p value was
obtained by comparing the moderate versus severe hydrocephalus genotype frequencies within each group. Neighboring SNPs rs6196597 and
rs6243755 also had similar genotype frequencies (data not shown)
Neurogenetics
Our data raise the possibility that genotypes at the candidate
chromosome 10 modifier may contribute to the observed
sex difference by contributing to moderate hydrocephalus
in females. However, the lack of correlation between 129
homozygosity at chromosome 10 and the phenotype of
severely affected females indicates that if the locus is
involved, it likely has a weak effect.
Expression analysis of genes in the L1hydro1 region
The L1hydro1 region, positioned between 100 and
129.2 Mb on chromosome 5, contains ∼565 known and/or
predicted Entrez listed genes (http://www.ncbi.nlm.nih.gov/
sites/entrez). While finer mapping will be required to
identify the gene(s) responsible for contributing to L1cam
X-linked hydrocephalus, we surveyed the expression
profiles of genes in the modifier region to gain insight
about possible candidates. We utilized two sets of microarray data. One was a publicly available GEO data set,
GDS1490, which included expression data from 24 neural
tissues in C57BL/6J and 129S6/SvEvTac (129S6) mice
[67], a substrain that like our 129S2 mice was derived from
the 129S1 129/Sv strain [69]. The second data set, obtained
in our lab, but previously unpublished, was an expression
profile of cerebellar tissue from postnatal day 11 congenic
129S2 L1cam knock-out mice and their wild-type siblings
(GEO data set: GSE13984).
Forty-three genes within the L1hydro1 region were
differentially expressed in neural tissues between C57BL/6J
and 129S6 mice by ≥1.5 fold (Supplemental Table 6).
RIKEN cDNA 2310001H12, a predicted gene containing a
Krüppel-associated (KRAB) box and a zinc finger domain,
demonstrated the greatest difference in expression. It was
upregulated 4.4-fold in the choroid plexus of 129S6 mice
compared to C57BL/6J mice. The choroid plexus is the
highly vascularized ependymal structure that lines the
ventricles of the brain and produces cerebrospinal fluid
(reviewed in [70]). Metal response transcription factor 2
(Mtf2) exhibited the second greatest difference in expression.
It was 3.81-fold higher in the choroid plexus of 129S6 versus
C57BL/6J mice. Mtf2 encodes a polycomb-like protein.
Polycomb group proteins silence transcription by binding
and methylating nucleosomes (reviewed in [71]). Interestingly, Mtf2 mutant mice exhibit hydrocephalus [8].
A few other genes that might play roles in L1cam function
and/or hydrocephalus were also differentially expressed. For
example, the genes encoding integrin binding secreted
phosphoprotein 1 (Spp1/osteopontin) ([72], reviewed in
[73]) and extracellular matrix-associated protein Sparc-like
1 (Sparcl1) [74] were upregulated in the choroid plexus of
129S6 versus C57BL/6J mice: 2.05- and 2.04-fold, respectively. Both Sparcl1 and Spp1, like L1cam, regulate cell
adhesion. In addition, polycystin 2 (Pkd2) was expressed
1.5-fold higher in the choroid plexus of 129S6 than C57BL/
6J mice. Pkd2, a member of the transient receptor potential
(TRP) channel family, can function as an intracellular
calcium release channel and as a cilium-anchored mechanosensory channel. In zebrafish, disruption of Pkd2 expression
caused hydrocephalus [21]. Aside from differential expression in the choroid plexus, several genes in the L1hydro1
region were also differentially expressed in other neural
tissues.
Analysis of the microarray study conducted by our lab
also provided some intriguing data. A total of three genes
located within the L1hydro1 region were differentially
expressed in cerebellum between 129S2 congenic L1KO
(129S2.Cg-L1camtm1Sor) and wild-type mice by at least
1.3-fold: Mtf2, Pkd2, and cell division cycle 7 (Cdc7),
which encodes a kinase involved in DNA replication [75].
All three genes were upregulated in wild-type mice, 1.79-,
1.36-, and 1.30-fold, respectively, compared to L1KOs. An
unpaired two-tailed Student’s t test was used to determine
significance. While the p values generated by Mtf2 and
Cdc7 (p<0.02) were significant, the p value for Pkd2 was
not (p=0.17). These results suggest a genetic interaction
between L1cam and genes located within the L1hydro1
region. Specifically, loss of L1cam seems to reduce the
expression of Mtf2 and Cdc7.
Discussion
A modifier locus on chromosome 5 contributes to L1cam
X-linked hydrocephalus
We propose that manifestation of L1cam X-linked hydrocephalus in mice is affected by the functional status of the
L1cam gene, as well as by the genotype of a modifier locus
on chromosome 5. As previously discussed, X-linked
hydrocephalus is evident in mice lacking all L1cam
function (L1KOs), as well as in L1-6D mice, which have
compromised L1-L1 homophilic and L1-integrin interactions. However, these mutations in the L1cam gene are not
sufficient for the development of severe hydrocephalus.
The phenotype is strain dependent. L1KO and L1-6D
mutant mice develop severe hydrocephalus when bred on
a C57BL/6J background. In contrast, 129S2 L1KO and L16D mutant mice do not. These data suggest that a locus,
which is polymorphic between the two strains, interacts
genetically with the L1cam gene to contribute to severe
hydrocephalus.
Our work advances the current understanding of how
hydrocephalus is caused by L1cam deficiencies by identifying a region on chromosome 5 (100–129.2 Mb) that
behaves as a L1cam hydrocephalus modifier locus in mice.
This locus, referred to as L1hydro1, surpassed Bonferroni
Neurogenetics
correction (α=0.005) and Lander and Krugylak’s significant linkage threshold in both the original genome-wide
scan and the extension study when analyzed with chisquare tests for deviations from Mendelian segregation.
Furthermore, our multiple QTL model, which was generated through a different set of statistical analyses, suggests
that the locus harbors a candidate severity modifier. Hence,
L1Hydro1 may affect both the severity of hydrocephalus in
L1cam mutants and the susceptibility of L1cam mutants to
the trait. As expected of a susceptibility modifier, a
significantly greater number of hydrocephalic F2 L1-6D
mutant mice were B6 homozygotes at L1Hydro1 than
predicted by Mendelian segregation. However, unaffected
B6 homozygotes were also observed, suggesting that the
modifier exhibits incomplete penetrance, a phenomenon
often observed when a trait is influenced by the interactions
of multiple loci.
Candidate regions for additional modifier loci
During the course of our analyses, we identified several
candidate modifier loci with weak linkage evidence. These
loci were found on chromosomes 2 (147.8–173.4 Mb), 3
(59.1–76.7 Mb), 4 (130.9–150.8), 7 (99.1 Mb), 9 (92.2–
102.2 Mb), and 10 (104.4–108.2 Mb). While these loci may
be false positives, it is also possible that some of these loci are
modifiers with small effects on L1cam X-linked hydrocephalus. Our data may therefore help guide future studies.
Since the loci on chromosomes 2 and 3 were detected by
chi-square tests for Mendelian segregation, they may affect
how susceptible L1cam mutants are to hydrocephalus. On
the other hand, the loci on chromosomes 7, 9, and 10 were
identified as candidate severity modifiers. QTL analysis
implicated the regions on chromosomes 7 and 9, while chisquare tests on groups segregated according to hydrocephalus severity detected the locus on chromosome 10. It is not
surprising that these distinct approaches identified different
candidate severity loci. While the stratified chi-square test
dichotomizes the hydrocephalus phenotype (moderate
versus severe), QTL analysis works with the full trait
distribution. Furthermore, while QTL analysis considers
gene–gene interactions, the stratified chi-square test can
only evaluate only how individual loci affect severity.
Interestingly, the locus on chromosome 4 was detected
by chi-square tests for Mendelian segregation, and it was
implicated by QTL analysis. It is possible that this locus,
like L1Hydro1, affects both susceptibility and severity.
Our findings suggest that X-linked hydrocephalus is a
complex trait. Genotypes at L1cam and L1hydro1 do not
completely explain all hydrocephalus cases in the F2
generation. Hence, it is likely that L1cam-associated Xlinked hydrocephalus is influenced by multiple genes. In
addition, since the L1-6D mutation perturbs L1 homophilic
and L1-integrin interactions [60], we suspect that disruption
of these types of interactions plays an important role in the
development of hydrocephalus.
Candidate modifier genes within the L1hydro1 region
Taken together, our linkage data, microarray analyses, and
published literature implicate Mtf2, as well as a few other
genes, including Pkd2, as strong candidates for the modifier
gene within L1hydro1. Strikingly, Mtf2 (108.5 Mb) and Pkd2
(104.9 Mb) even flank the chromosome 5 candidate severity
QTL (106.7 Mb). In addition, Mtf2 microarray results
illustrate how the expression of the modifier gene in L1hydro1
may be affected by genetic background. Data suggest that
C57BL/6J mice express less Mtf2 than 129S2 mice. Furthermore, L1cam knock-out mice exhibited decreased Mtf2
expression compared to wild-type mice. Following this
example, hydrocephalus may be prevalent in C57BL/6J but
not 129/Sv L1cam mutants for the following reason: (1)
L1cam mutants may express the modifier gene at lower levels
than wild-type mice, and (2) C57BL/6J mice may exhibit
decreased modifier gene expression compared to 129S2 mice.
Nevertheless, we also performed a more traditional search,
utilizing NCBI databases such as PubMed and Entrez Gene, to
search for additional candidate modifier genes within the
L1hydro1 region. Fibroblast growth factor receptor like-1
(Fgfrl1) emerged as an intriguing candidate because L1 has
been shown to directly bind FGFR [38]. Mapk10 (JNK3) is
also an interesting candidate because pharmacological
studies suggest that inhibition of the JNK pathway affects
L1 function (unpublished results). Furthermore, we consider
transforming growth factor beta receptor III (Tgfβr3) a
strong candidate because the receptor binds TGFβ1 [76], as
well as other TGFβ superfamily members [77]. Transgenic
mice that overproduce active Tgfβ1 in the CNS develop
hydrocephalus [14]. Lastly, two dynein subunits in the
region, Dynein light chain LC8-type 1 (Dynll1) and Dynein
axonemal heavy chain 10 (Dnahc10) stood out because loss
of function of a related gene, Dynein axonemal heavy chain
5 (Dnahc5), results in hydrocephalus [20].
Future work will aim to narrow down the L1hydro1 region
and identify the modifier gene. The creation of L1-6D 129S2
congenic mice that are B6 homozygotes within the L1hydro1
interval should facilitate mapping and gene identification, as
mice that carry the B6 allele at the modifier locus are
expected to exhibit hydrocephalus. On the other hand,
sequencing candidates may not be very informative. For
example, sequence information from the Mouse Phenome
Database indicates that several of the candidate genes
discussed above contain nonsynonymous SNPs in their
coding regions, yet, whether these polymorphisms contribute
to hydrocephalus cannot be known without more functional
experiments.
Neurogenetics
Human homologs in the L1hydro1 region
The genes within the L1hydro1 region have 297 human
homologs (www.ensembl.org). Syntenic tracts of the
L1hydro1 region are dispersed throughout several human
chromosomes. Most of the L1hydro1 region maps to
human chromosomes 1, 4, 12, and 22. However, there are
also a small percentage of genes on the human X
chromosome as well as on chromosomes 5, 11, and 16.
Our search through the Pubmed and OMIM databases
revealed no genes homologous to those in the L1hydro1
region that have previously been associated with human
hydrocephalus.
In conclusion, the identification of the L1hydro1 locus
and additional candidate modifier loci has driven the field
closer to characterizing genes that contribute to hydrocephalus and genetically interact with L1cam. The insights
obtained from this line of study will help determine whether
genes that influence the development of congenital hydrocephalus in the mouse have similar roles in humans. The
discovery of L1cam modifier genes will enhance our
understanding of the molecular and cellular pathology of
congenital hydrocephalus, a complex trait that develops
from the interplay of multiple genetic factors. Furthermore,
by elucidating the relationships between L1cam and its
modifier genes, we will better comprehend L1cam signaling
and function. Indeed, in the past, modifier gene studies
have made vital contributions to our understanding of
signal transduction mechanisms, and they have revealed
novel pathway components [78].
Acknowledgments We thank Shrikant Mane and Sheila Westman
from the Yale Microarray Center, Jennifer Moran from the Mouse
SNP Genotyping Service, Mutation Mapping and Developmental
Analysis Project at Brigham and Women’s Hospital, Genetics
Division, Harvard Medical School, and Alison Brown from the
Harvard Partners Center for Genetics and Genomics for SNP
genotyping and evaluation. We also thank the Case Western Reserve
Transgenic and Targeting Facility for their in vitro fertilization testing
of L1KO mice. Our nomenclature for the L1hydro1 locus was
approved by the Mouse Genomic Nomenclature Committee of the
Mouse Genome Informatics Resource, Jackson Laboratory. This work
was supported by the Miami Project to Cure Paralysis and by National
Institute of Health grants HD39884, EY05285, N01-NS-3-2351 and
T32NS07459. V. Lemmon holds the Walter G. Ross Distinguished
Chair in Developmental Neuroscience at the University of Miami.
References
1. Network NBDP (2007) Congenital malformations surveillance
report: Birth defects surveillance data from selected states, 2000–
2004. Birth Defects Research Part A 76:874–942
2. Persson EK, Anderson S, Wiklund LM, Uvebrant P (2007)
Hydrocephalus in children born in 1999–2002: epidemiology,
outcome and ophthalmological findings. Childs Nerv Syst
23:1111–1118
3. Jouet M, Rosenthal A, MacFarlane J, Kenwrick S, Donnai D
(1993) A missense mutation confirms the L1 defect in X-linked
hydrocephalus (HSAS). Nat Genet 4:331
4. Halliday J, Chow CW, Wallace D, Danks DM (1986) X linked
hydrocephalus: a survey of a 20 year period in Victoria, Australia.
J Med Genet 23:23–31
5. Vincent C, Kalatzis V, Compain S, Levilliers J, Slim R et al
(1994) A proposed new contiguous gene syndrome on 8q consists
of Branchio-Oto-Renal (BOR) syndrome, Duane syndrome, a
dominant form of hydrocephalus and trapeze aplasia; implications
for the mapping of the BOR gene. Hum Mol Genet 3:1859–1866
6. Dahme M, Bartsch U, Martini R, Anliker B, Schachner M et al
(1997) Disruption of the mouse L1 gene leads to malformations of
the nervous system. Nat Genet 17:346–349
7. Cohen NR, Taylor JS, Scott LB, Guillery RW, Soriano P et al
(1998) Errors in corticospinal axon guidance in mice lacking the
neural cell adhesion molecule L1. Curr Biol 8:26–33
8. Wang S, He F, Xiong W, Gu S, Liu H et al (2007) Polycomblike2-deficient mice exhibit normal left-right asymmetry. Dev Dyn
236:853–861
9. Hirotsune S, Fleck MW, Gambello MJ, Bix GJ, Chen A et al
(1998) Graded reduction of Pafah1b1 (Lis1) activity results in
neuronal migration defects and early embryonic lethality. Nat
Genet 19:333–339
10. Hong HK, Chakravarti A, Takahashi JS (2004) The gene for
soluble N-ethylmaleimide sensitive factor attachment protein
alpha is mutated in hydrocephaly with hop gait (hyh) mice. Proc
Natl Acad Sci U S A 101:1748–1753
11. Chae TH, Kim S, Marz KE, Hanson PI, Walsh CA (2004) The
hyh mutation uncovers roles for alpha Snap in apical protein
localization and control of neural cell fate. Nat Genet 36:264–270
12. das Neves L, Duchala CS, Tolentino-Silva F, Haxhiu MA,
Colmenares C et al (1999) Disruption of the murine nuclear
factor I-A gene (Nfia) results in perinatal lethality, hydrocephalus,
and agenesis of the corpus callosum. Proc Natl Acad Sci U S A
96:11946–11951
13. Zhang J, Williams MA, Rigamonti D (2006) Genetics of human
hydrocephalus. J Neurol 253:1255–1266
14. Galbreath E, Kim SJ, Park K, Brenner M, Messing A (1995)
Overexpression of TGF-beta 1 in the central nervous system of
transgenic mice results in hydrocephalus. J Neuropathol Exp
Neurol 54:339–349
15. Wyss-Coray T, Feng L, Masliah E, Ruppe MD, Lee HS et al
(1995) Increased central nervous system production of extracellular matrix components and development of hydrocephalus in
transgenic mice overexpressing transforming growth factor-beta 1.
Am J Pathol 147:53–67
16. Zygourakis CC, Rosen GD (2003) Quantitative trait loci modulate
ventricular size in the mouse brain. J Comp Neurol 461:362–369
17. Jones HC, Yehia B, Chen GF, Carter BJ (2004) Genetic analysis
of inherited hydrocephalus in a rat model. Exp Neurol 190:79–90
18. Davy BE, Robinson ML (2003) Congenital hydrocephalus in hy3
mice is caused by a frameshift mutation in Hydin, a large novel
gene. Hum Mol Genet 12:1163–1170
19. Lechtreck KF, Delmotte P, Robinson ML, Sanderson MJ, Witman
GB (2008) Mutations in Hydin impair ciliary motility in mice. J
Cell Biol 180:633–643
20. Ibanez-Tallon I, Gorokhova S, Heintz N (2002) Loss of function
of axonemal dynein Mdnah5 causes primary ciliary dyskinesia
and hydrocephalus. Hum Mol Genet 11:715–721
21. Obara T, Mangos S, Liu Y, Zhao J, Wiessner S et al (2006)
Polycystin-2 immunolocalization and function in zebrafish. J Am
Soc Nephrol 17:2706–2718
22. Pathak N, Obara T, Mangos S, Liu Y, Drummond IA (2007) The
zebrafish fleer gene encodes an essential regulator of cilia tubulin
polyglutamylation. Mol Biol Cell 18:4353–4364
Neurogenetics
23. Ferrante MI, Romio L, Castro S, Collins JE, Goulding DA et al
(2009) Convergent extension movements and ciliary function are
mediated by ofd1, a zebrafish orthologue of the human oral-facialdigital type 1 syndrome gene. Hum Mol Genet 18:289–303
24. Takeda Y, Asou H, Murakami Y, Miura M, Kobayashi M et al
(1996) A nonneuronal isoform of cell adhesion molecule L1:
tissue-specific expression and functional analysis. J Neurochem
66:2338–2349
25. Kamiguchi H, Lemmon V (1997) Neural cell adhesion molecule
L1: signaling pathways and growth cone motility. J Neurosci Res
49:1–8
26. Esch T, Lemmon V, Banker G (2000) Differential effects of
NgCAM and N-cadherin on the development of axons and
dendrites by cultured hippocampal neurons. J Neurocytol
29:215–223
27. Lagenaur C, Lemmon V (1987) An L1-like molecule, the 8D9
antigen, is a potent substrate for neurite extension. Proc Natl Acad
Sci U S A 84:7753–7757
28. Kunz S, Spirig M, Ginsburg C, Buchstaller A, Berger P et al
(1998) Neurite fasciculation mediated by complexes of axonin-1
and Ng cell adhesion molecule. J Cell Biol 143:1673–1690
29. Stallcup WB, Beasley L (1985) Involvement of the nerve growth
factor-inducible large external glycoprotein (NILE) in neurite
fasciculation in primary cultures of rat brain. Proc Natl Acad Sci
U S A 82:1276–1280
30. Itoh K, Fushiki S, Kamiguchi H, Arnold B, Altevogt P et al
(2005) Disrupted Schwann cell-axon interactions in peripheral
nerves of mice with altered L1-integrin interactions. Mol Cell
Neurosci 30:624–629
31. Lindner J, Rathjen FG, Schachner M (1983) L1 mono- and
polyclonal antibodies modify cell migration in early postnatal
mouse cerebellum. Nature 305:427–430
32. Itoh K, Shimono K, Lemmon V (2005) Dephosphorylation and
internalization of cell adhesion molecule L1 induced by theta
burst stimulation in rat hippocampus. Mol Cell Neurosci 29:245–
249
33. Nakamura Y, Tamura H, Horinouchi K, Shiosaka S (2006) Role of
neuropsin in formation and maturation of Schaffer-collateral
L1cam-immunoreactive synaptic boutons. J Cell Sci 119:1341–
1349
34. Triana-Baltzer GB, Liu Z, Berg DK (2006) Pre- and postsynaptic
actions of L1-CAM in nicotinic pathways. Mol Cell Neurosci
33:214–226
35. Lemmon V, Farr KL, Lagenaur C (1989) L1-mediated axon
outgrowth occurs via a homophilic binding mechanism. Neuron
2:1597–1603
36. Castellani V, De Angelis E, Kenwrick S, Rougon G (2002) Cis
and trans interactions of L1 with neuropilin-1 control axonal
responses to semaphorin 3A. Embo J 21:6348–6357
37. Doherty P, Walsh FS (1996) CAM-FGF receptor interactions: a
model for axonal growth. Mol Cell Neurosci 8:99–111
38. Kulahin N, Li S, Hinsby A, Kiselyov V, Berezin V et al (2008)
Fibronectin type III (FN3) modules of the neuronal cell adhesion
molecule L1 interact directly with the fibroblast growth factor
(FGF) receptor. Mol Cell Neurosci 37:528–536
39. Schmid RS, Pruitt WM, Maness PF (2000) A MAP kinasesignaling pathway mediates neurite outgrowth on L1 and requires
Src-dependent endocytosis. J Neurosci 20:4177–4188
40. Schaefer AW, Kamei Y, Kamiguchi H, Wong EV, Rapoport I et al
(2002) L1 endocytosis is controlled by a phosphorylationdephosphorylation cycle stimulated by outside-in signaling by
L1. J Cell Biol 157:1223–1232
41. Schaefer AW, Kamiguchi H, Wong EV, Beach CM, Landreth G et
al (1999) Activation of the MAPK signal cascade by the neural
cell adhesion molecule L1 requires L1 internalization. J Biol
Chem 274:37965–37973
42. Dickson TC, Mintz CD, Benson DL, Salton SR (2002) Functional
binding interaction identified between the axonal CAM L1 and
members of the ERM family. J Cell Biol 157:1105–1112
43. Cheng L, Itoh K, Lemmon V (2005) L1-mediated branching is
regulated by two ezrin-radixin-moesin (ERM)-binding sites, the
RSLE region and a novel juxtamembrane ERM-binding region. J
Neurosci 25:395–403
44. Davis JQ, Bennett V (1994) Ankyrin binding activity shared by
the neurofascin/L1/NrCAM family of nervous system cell
adhesion molecules. J Biol Chem 269:27163–27166
45. Hortsch M, Nagaraj K, Godenschwege TA (2009) The interaction
between L1-type proteins and ankyrins—a master switch for L1type CAM function. Cell Mol Biol Lett 14:57–69
46. Cheng L, Lemmon S, Lemmon V (2005) RanBPM is an L1interacting protein that regulates L1-mediated mitogen-activated
protein kinase activation. J Neurochem 94:1102–1110
47. Yamasaki M, Thompson P, Lemmon V (1997) CRASH syndrome:
mutations in L1CAM correlate with severity of the disease.
Neuropediatrics 28:175–178
48. Kamiguchi H, Hlavin ML, Yamasaki M, Lemmon V (1998)
Adhesion molecules and inherited diseases of the human nervous
system. Annu Rev Neurosci 21:97–125
49. Fransen E, Van Camp G, D'Hooge R, Vits L, Willems PJ (1998)
Genotype-phenotype correlation in L1 associated diseases. J Med
Genet 35:399–404
50. Schrander-Stumpel C, Howeler C, Jones M, Sommer A, Stevens
C et al (1995) Spectrum of X-linked hydrocephalus (HSAS),
MASA syndrome, and complicated spastic paraplegia (SPG1):
clinical review with six additional families. Am J Med Genet
57:107–116
51. Willems PJ, Brouwer OF, Dijkstra I, Wilmink J (1987) X-linked
hydrocephalus. Am J Med Genet 27:921–928
52. Yamasaki M, Arita N, Hiraga S, Izumoto S, Morimoto K et al
(1995) A clinical and neuroradiological study of X-linked
hydrocephalus in Japan. J Neurosurg 83:50–55
53. Renier WO, Ter Haar BG, Slooff JL, Hustinx TW, Gabreels FJ
(1982) X-linked congenital hydrocephalus. Clin Neurol Neurosurg
84:113–123
54. Boyd E, Schwartz CE, Schroer RJ, May MM, Shapiro SD et al
(1993) Agenesis of the corpus callosum associated with MASA
syndrome. Clin Dysmorphol 2:332–341
55. Fried K (1972) X-linked mental retardation and-or hydrocephalus.
Clin Genet 3:258–263
56. Kaepernick L, Legius E, Higgins J, Kapur S (1994) Clinical
aspects of the MASA syndrome in a large family, including
expressing females. Clin Genet 45:181–185
57. Serville F, Lyonnet S, Pelet A, Reynaud M, Louail C et al (1992)
X-linked hydrocephalus: clinical heterogeneity at a single gene
locus. Eur J Pediatr 151:515–518
58. Fransen E, D'Hooge R, Van Camp G, Verhoye M, Sijbers J et al
(1998) L1 knockout mice show dilated ventricles, vermis
hypoplasia and impaired exploration patterns. Hum Mol Genet
7:999–1009
59. Rolf B, Kutsche M, Bartsch U (2001) Severe hydrocephalus in
L1-deficient mice. Brain Res 891:247–252
60. Itoh K, Cheng L, Kamei Y, Fushiki S, Kamiguchi H et al (2004)
Brain development in mice lacking L1-L1 homophilic adhesion. J
Cell Biol 165:145–154
61. Hof PR, Young WG, Bloom FE, Belichenko PV, Celio MR (2000)
Comparative cytoarchitectonic atlas of the C57BL/6 and 129/Sv
mouse brains. Elsevier, New York
62. Abramoff MD, Magelhaes PJ, Ram SJ (2004) Image processing
with ImageJ. Biophoton Int 11:36–42
63. Lorentzen JC, Glaser A, Jacobsson L, Galli J, Fakhrai-rad H et al
(1998) Identification of rat susceptibility loci for adjuvant-oilinduced arthritis. Proc Natl Acad Sci U S A 95:6383–6387
Neurogenetics
64. Varga L, Muller G, Szabo G, Pinke O, Korom E et al (2003)
Mapping modifiers affecting muscularity of the myostatin mutant
(Mstn(Cmpt-dl1Abc)) compact mouse. Genetics 165:257–267
65. Broman KW, Wu H, Sen S, Churchill GA (2003) R/qtl: QTL
mapping in experimental crosses. Bioinformatics 19:889–890
66. Broman KW, Sen S, Owens SE, Manichaikul A, Southard-Smith
EM et al (2006) The X chromosome in quantitative trait locus
mapping. Genetics 174:2151–2158
67. Zapala MA, Hovatta I, Ellison JA, Wodicka L, Del Rio JA et al
(2005) Adult mouse brain gene expression patterns bear an
embryologic imprint. Proc Natl Acad Sci U S A 102:10357–
10362
68. Lander E, Kruglyak L (1995) Genetic dissection of complex traits:
guidelines for interpreting and reporting linkage results. Nat Genet
11:241–247
69. Beck JA, Lloyd S, Hafezparast M, Lennon-Pierce M, Eppig JT et
al (2000) Genealogies of mouse inbred strains. Nat Genet 24:23–
25
70. Redzic ZB, Segal MB (2004) The structure of the choroid plexus
and the physiology of the choroid plexus epithelium. Adv Drug
Deliv Rev 56:1695–1716
71. Schwartz YB, Pirrotta V (2007) Polycomb silencing mechanisms and
the management of genomic programmes. Nat Rev, Genet 8:9–22
72. Hu DD, Lin EC, Kovach NL, Hoyer JR, Smith JW (1995) A
biochemical characterization of the binding of osteopontin to integrins
alpha v beta 1 and alpha v beta 5. J Biol Chem 270:26232–26238
73. Wai PY, Kuo PC (2008) Osteopontin: regulation in tumor
metastasis. Cancer Metastasis Rev 27:103–118
74. Gongidi V, Ring C, Moody M, Brekken R, Sage EH et al (2004)
SPARC-like 1 regulates the terminal phase of radial glia-guided
migration in the cerebral cortex. Neuron 41:57–69
75. Kim JM, Masai H (2004) Genetic dissection of mammalian Cdc7
kinase: cell cycle and developmental roles. Cell Cycle 3:300–304
76. Andres JL, Ronnstrand L, Cheifetz S, Massague J (1991)
Purification of the transforming growth factor-beta (TGF-beta)
binding proteoglycan betaglycan. J Biol Chem 266:23282–23287
77. Kirkbride KC, Townsend TA, Bruinsma MW, Barnett JV, Blobe
GC (2008) Bone morphogenetic proteins signal through the
transforming growth factor-beta type III receptor. J Biol Chem
283:7628–7637
78. St Johnston D (2002) The art and design of genetic screens:
drosophila melanogaster. Nat Rev, Genet 3:176–188